|Abstract:||A localization system allows a robot to determine its location. Many algorithms are available
with different tradeoffs. This research addresses the implementation of a positioning algorithm on
ultra-wideband (UWB) transceivers for indoor positioning. There are various positioning
systems suitable for indoor robotics, which can be categorized as signal-strength-based and
time-based. Some of the time-based algorithms are time of flight (ToA), time difference of
arrival (TDoA), and angle of arrival (AoA).
In this paper we focus on the Sequential Time Difference of Arrival (S-TDoA) algorithm. Unlike
TDoA, S-TDoA does not require clock synchronization. This method also does not transmit
signals that ToA needs, which provides several benefits, one being scalability as the update rate
is irrelevant to the number of tags, the other being privacy preservation.
This thesis develops statistical analysis on temporal and spatial data to improve performance.
We devise several metrics to measure various positioning algorithm performance, all in
comparison with the Vicon system where we assume it to be the ground truth. The preliminary
result shows, for each position estimate, the mean offset error around 0.1 meters on x- and y-axes
with the z-axis offset error of 0.7 meters. The mean-variance for the estimate is 0.002 meters on the
x- and y-axes and 0.014 meters on the z-axis.